Screening of Candidate Housekeeping Genes in Uterus Caruncle by RNA-Sequence and qPCR Analyses in Different Stages of Goat (Capra hircus)

Simple Summary

The uterus is an important organ in the reproductive system of most female mammals. The normal growth and development of ruminant uterus caruncles are crucial to maintain gestation and fetal health in goats. The aim of this study was to screen a set of candidate housekeeping genes (HKGs) in Carpra hircus uterus caruncles. Four programs and one comprehensive index were applied to assess the stability of 22 selected HKGs expression during non-pregnant and pregnancy stages. The most and least stable HKGs were used to normalize the target genes expression of SPP1, VEGFA, and PAG8. Traditional reference genes were not suitable for target gene normalization, while PPIB and EIF3K, POP4, PPIA, and SDHA showed the least variation and were recommended as the best HKGs during the nonpregnant stage and the whole stages of goat uterus caruncle tissue, respectively. This study found the suitable HKGs in uterus caruncle tissues at different stages of non-pregnancy and pregnancy.

Abstract

The uterus is a critical pregnancy organ for mammals. The normal growth and development of ruminant uterus caruncles are crucial to maintain gestation and fetal health in goats. Quantitative real-time polymerase chain reaction (qRT-PCR) is a reliable tool to study gene expression profiling for exploring the intrinsic mechanism underlying the conversion process of uterus caruncle tissue. However, the candidate housekeeping genes (HKGs) are required for normalizing the expression of function genes. In our study, 22 HKGs were selected from analyzing transcriptome data at non-pregnancy and pregnancy processes and previous reports about HKGs in goat tissues. We assessed them for expression suitability in 24 samples from uterus tissues at 15 non-pregnant days (Stage 1), early (Stage 2), and medium-later pregnant days (Stage 3). The expression stability of these genes was evaluated by using geNorm, Normfinder, Bestkeeper, and Delta Ct algorithms and, comprehensively, by ReFinder. In addition, the most and least stable HKGs were used to normalize the target genes expression of SPP1, VEGFA, and PAG8. It was found that traditional reference genes, such as ACTB and GAPDH, were not suitable for target gene normalization. In contrast, PPIB selected from RNA sequencing data and EIF3K selected from previous references showed the least variation and were recommended as the best HKGs during the nonpregnant stage and the whole stages of goat uterus caruncle tissue, respectively. It is the first time the HKGs genes in uterus during the non-pregnant day and throughout the total pregnancy have been explored. These findings found suitable HKGs in uterus caruncle tissues at various stages of non-pregnancy and pregnancy; these can be useful for gene expression studies to reveal the molecular mechanisms of uterus development in goats.

1. Introduction

The uterus, a place where the embryo is implanted and continues to develop, is necessary for the reproduction of mammals. Uterus caruncles are a special structure located on the endometrium of domestic ruminants. During the process of pregnancy, uterus caruncles participate in placentation and form to become the maternal part of placenta [1]. Before and after pregnancy, the uterus caruncle develops from endometrial epithelial cells to connective tissue rich in blood vessels. This dynamic process indicates that the structure and cell types of the uterus caruncle undergo significant changes during the pregnancy.

Goat (Capra hircus) is becoming widely accepted around the world due to the increasing demand for the low cholesterol content and high nutritive value of goat meat [2]. Successful pregnancy and fetal health in goats are important safeguards for reproductive performance and are dependent on normal uterus function and placenta development. The hircine placenta is the cotyledon placenta, which divides into multiple button-shaped functional units, entrenched into maternal uterus caruncle with the incomplete uterine luminal epithelium, and achieves the synepitheliochorial placenta at gestation [3,4]. Although several physiological processes have been reported in ruminant uteruses [5,6], the exact molecular mechanisms remain largely unclear in uterus caruncle development in domestic animals. Gene expression analysis is a useful technique because it provides information on the regulation of uterus development.

Quantitative real-time polymerase chain reaction (qRT-PCR) is widely applied to quickly and accurately characterize the expression of target genes. To obtain the real expression changes of target genes, HKGs must be used to normalize cycle threshold (Ct) value from qRT-PCR. In theory, the expression of HKGs should not change in the investigated different tissues or in development stages. To obtain the candidate housekeeping genes, programs (including geNorm [7], Normfinder [8], BestKeeper [9], and Delta-Ct method [10]) have been invented to facilitate the statistical analysis of potential endogenous control genes base on the results of qRT-PCR. Furthermore, comprehensive tools have also been developed to compare tested HKGs, including RefFinder [11] and ComprFinder [12]. Meanwhile, high-throughput RNA sequencing provided a large transcriptional collection of all genes, and the fragments per kilobase of transcript sequence per millions of base pairs sequenced (PFKM) has been used to screen for new or stable reference genes with low coefficient variation [13,14].

To date, several studies have identified the candidate HKGs in uteruses of other species, but not goats. For instance, the combination of YWHAZ, HPRT1, and HMBS was suggested as the most stable reference group of genes in the ovarian and uterine tissues of laying hens under control and heat stress conditions [15]; YWHAZ was the best reference gene, which could be used as an accurate internal control gene in canine pyometra studies [16]. Moreover, there were also some studies carried out on tissues other than the uterus of goat. The HMBS and B2M were identified as the candidate housekeeping genes in hircine spleen, caecum, small intestine, lower lip, large intestine, trachea, and upper lip [14]. Several HKGs were also demonstrated in other tissues of goats, such as SDHA in skin [12], EIF3M in adipose tissue [13], and PPIB and HMBS in intramuscular fat [17]. However, the candidate housekeeping genes for use in the study of goat uteruses, especially uterus caruncle, are lacking.

In this study, we aimed to screen a set of candidate HKGs in Carpra hircus uterus caruncle. Twenty-two original candidate HKGs were selected by analyzing transcriptome data and referring to previous studies. The transcriptional expression levels of these genes were detected in uterus caruncle by qRT-PCR. Four programs and one comprehensive index were applied to assess the stability of HKGs expression during non-pregnant and pregnancy stages. These findings were useful in selecting HKGs to correctly analyze the expression of function genes in the uterus caruncle, which aids in understanding the molecular genetic mechanism of pregnancy in goats.

2. Materials and Methods

2.1. Ethics Statement

All goat experiments were performed in strict accordance with the recommendations of the Southwest University Institutional Animal Care and Use Committee (2019, No. GB14925-2010). All research involving animals was conducted according to the approved protocols of the Chongqing Key Laboratory of Herbivore Science, Chongqing, China.

2.2. Animals and Sample Harvesting

The research involving animals was conducted according to the approved protocols of the Chongqing Key Laboratory of Herbivore Science, Chongqing, China. The animals were from the breeding farm at Southwest University. They were subjected to a standard feeding management regime (DB50/T 510-2013) and housed in free semi-open folds. Sample procedures in this study were supervised and approved by the Institutional Animal Care and Use Committee of Southwest University. Through cesarean operation, uterus caruncle samples were collected at eight different non-pregnant and pregnant stages of Dazu Black goats (Figure 1a,b).

2.3. RNA Extraction and cDNA Synthesis

Total RNA is extracted from uterus caruncle tissues using Trizol^® reagent according to the manufacturer’s instructions (Invitrogen, Carlsbad, CA, USA). RNA concentration was measured by NanoDrop 2000 spectrophotometer (Thermo Fisher, Meridian, ID, USA), and the purity of RNA was determined by the ratio of OD260/280. All the ratio of OD260/280 of our RNA samples value 1.8–2.0. Each sample was reverse transcribed using All-In-One 5X RT MasterMix (Abm, Heidelberg, Germany, Cat#G592) with 2 μg total RNA.

2.4. Selection of Candidate Reference Genes

Transcriptome sequencing data of 9 Dazu Black goat uterus caruncle tissues were obtained using the paired-end sequencing technology on an Illumina NovaSeq 6000 platform. After assembly and annotation, the gene expression profiles and read counts of unique transcripts were converted into FPKM values on the platform BMKCloud (www.biocloud.net) (accessed on 15 March 2021), according to the formula FPKM = cDNA fragments/[mapped fragments (millions) × transcript length (kb)]. Based on the FPKM value, every gene in each transcriptome, the CV, and MFC were calculated using Microsoft Excel. The CV is defined as the ratio of the SDFPKM to the mean of the FPKM of all samples for one gene. The MFC, which is defined as the fold change between the largest and smallest FPKM values within 9 RNA-seq, was calculated. A potential HKG should have a relatively high FPKM value [12,18] and low CV and MFC values because these are the typical requirements for HKGs, along with a relatively high and low expression variation level. We also got several HKGs from prior studies [19,20,21] and HKGs of frequently used genes in addition to genes chosen from transcriptome sequencing data with ideal FPKM, CV, and MFC values. qRT-PCR was used to amplify all HKGs for later analysis and confirmation.

2.5. Quantitative Real-Time PCR (qRT-PCR) Analysis

Specific primers were designed using the Primer-Blast web tool (https://www.ncbi.nlm.nih.gov/tools/primer-blast/ accessed on 22 March 2021) based on the sequences of unigenes. The information on primers is shown in Table S1. The qRT-PCR was carried out in a CFX96 Real-Time System (BIO-RAD, Hercules, CA, USA) with SsoAdvanced SYBR Green Supermix (BIO-RAD, #1725262). The qRT-PCR reaction system was performed in a 10 μL reaction volume with 5 μL of 2× qPCR MasterMix (Abm, Richmond, BC, Canada, Cat#G891), 4.5 μL cDNA template, and 0.25 μL each primer. The thermocycling conditions were performed starting with a 3 min enzyme activation step at 95 °C, followed by 40 cycles of 15 s at 95 °C for denaturation, and 1 min at 60 °C for annealing/extension. Melting curve analysis was performed at the end of each PCR program to monitor non-specific product formation.

2.6. Analysis of the Expression Stability for Candidate Reference Genes

The qRT-PCR Ct (Cycle Threshold) data were entered into four traditional algorithms, including geNorm [7], NormFinder [8], BestKeeper [9], and the Delta Ct method [10], to evaluate the stability of HKGs. Moreover, ReFinder [11] was used to obtain a comprehensive ranking of HKGs.

2.7. Validation of HKGs

To validate the HKGs stability results, the expression profiles of 3 well known goat uterus caruncle genes (VEGFA, PAG8, and SPP1) were investigated by performing qRT-PCR analysis of cDNA samples, in the same way as stated above. The expression levels of target genes were normalized using the 2^−∆∆CT method [22]. The primer details of these target caruncle genes have been provided in the Supplementary Table S1.

3. Results

3.1. Selection of HKGs Using RNA-Seq Data

Based on the FPKM value, the coefficient of variation (CV, %), and the maximum fold change (MFC), we obtained a total of 383 HKGs (Figure 2). KEGG (Kyoto Encyclopedia of Genes and Genomes) enrichment analysis indicated that most HKGs were involved in the ribosome and lysosome pathways (Figure S1). Five HKGs, including MORF4L1, HNRNPL, PTPRA, ARF1, and PPIB, were selected for further validation. We acquired several HKGs from prior studies [20] and HKGs of widely used genes (GAPDH, SDHA, ACTB, YWHAZ, and 18S) in addition to genes chosen from transcriptome sequencing data.

Table 1. The information of selected candidate housekeeping genes in RNA-seq data of caruncle at non-pregnant and early pregnant days.

Gene Sources	Rank	Mean FPKM	CV	MFC	Symbol	Description
Transcriptome Data	1	378.38	0.03	1.11	MORF4L1	mortality factor 4 like 1 [Source:NCBI gene;Acc:102179420]
	2	190.26	0.04	1.15	ARF1	ADP ribosylation factor 1 [Source:NCBI gene;Acc:102185451]
	4	171.61	0.05	1.17	HNRNPL	heterogeneous nuclear ribonucleoprotein L [Source:NCBI gene;Acc:102180618]
	9	50.26	0.05	1.22	PTPRA	protein tyrosine phosphatase receptor type A [Source:NCBI gene;Acc:102184258]
	20	89.19	0.06	1.22	PPIB	peptidylprolyl isomerase B [Source:NCBI gene;Acc:102169391]

contains all the details of the HKGs we obtained from transcriptome data for our research.

3.2. Amplification Specificity and Efficiency of the HKGs and Target Genes

A total of 25 primer pairs, including 22 HKGs and three target genes, were designed for qRT-PCR experiment. The specificity for each paired primer was validated by the melting curve analysis, which showed a single amplification peak (Figure S2). Each pair of primers had good specificity and amplification efficiency around 100%.

3.3. Expression Profiling of the HKGs

The expression of 22 HKGs was assayed in a sample of caruncle. The distribution of the mean Ct values (the average of three technical replicates in the same sample) is displayed in Figure 3a. The mean Ct values range from 19.80 to 32.87. We selected genes with a mean Ct value lower than 28 for our subsequent analysis. In other words, HKGs whose mean Ct value was greater than 28, including POLR2A, TBP, PTPRA, 18S, and UBC, were excluded in further analysis. These findings suggest that gene expression varied from approximately 300 folds per change compared to the mean gene expression of all 17 genes (Figure 3b).

3.4. Stability Analysis of HKGs by Four Algorithms

3.4.1. Expression Stability of the HKGs by Delta Ct Analysis

Using the Delta Ct method, gene stability was evaluated by the std-value, shown in

Table 2. Expression stability of HKGs by Delta CT analysis.

	Stage1		Stage2		Stage3		Total
Rank	Gene	STDEV	Gene	STDEV	Gene	STDEV	Gene	STDEV
1	PPIB	0.57	EIF3K	1.67	EIF3K	1.14	EIF3K	1.47
2	POP4	0.59	UXT1	1.73	PPIA	1.17	POP4	1.61
3	YWHAZ	0.61	EIF3M	1.85	POP4	1.21	SDHA	1.62
4	HMBS	0.61	MORF4L1	1.88	GAPDH	1.21	PPIA	1.67
5	MORF4L1	0.62	HPRT1	1.94	SDHA	1.23	EIF3M	1.70
6	HPRT1	0.63	HMBS	1.94	HMBS	1.25	PPIB	1.71
7	EIF3K	0.67	PPIB	1.97	EIF3M	1.25	HMBS	1.72
8	SDHA	0.68	SDHA	1.99	HNRNPL	1.30	GAPDH	1.75
9	EIF3M	0.68	POP4	1.99	ACTB	1.32	HNRNPL	1.77
10	RPLPO	0.69	PPIA	2.10	YWHAZ	1.35	UXT1	1.79
11	GAPDH	0.74	B2M	2.14	UXT1	1.37	YWHAZ	1.84
12	HNRNPL	0.76	GAPDH	2.22	PPIB	1.40	MORF4L1	1.87
13	UXT1	0.86	HNRNPL	2.23	ARF1	1.46	HPRT1	1.88
14	PPIA	0.89	YWHAZ	2.34	HPRT1	1.55	ARF1	2.17
15	ARF1	1.01	ACTB	2.36	B2M	1.81	ACTB	2.23
16	B2M	1.19	RPLPO	2.39	MORF4L1	1.82	B2M	2.48
17	ACTB	1.23	ARF1	2.94	RPLPO	3.35	RPLPO	2.91

. At stage 1, PPIB, a HKG we selected from RNA-seq data, demonstrated the best stability. However, it failed to be the most stable gene in other stages of all uterus caruncle samples, which EIF3K proved to be. Additionally, the other three HKGs selected from RNA-seq data (ie, MORF4L1, HNRNPL, and ARF1) did not demonstrate good stability.

3.4.2. Expression Stability of the HKGs by BestKeeper Analysis

The BestKeeper algorithm determines the best HKGs based on the principle that the expression of good HKGs are highly correlated and constructs a correlative index from the results, which is then compared with the target genes to reach a conclusion whether or not the reference genes are regulated. EIF3K was predicted to be the most stable among stage 2, stage 3, and overall samples by the BestKeeper algorithm, with the stability value (standard deviation) of 0.46, 0.36, and 0.46, respectively (

Table 3. Expression stability of HKGs by BestKeeper analysis.

	Stage1			Stage2			Stage3			Total
Rank	Gene	Std Dev	Coeff. of Corr. [r]	Gene	Std Dev	Coeff. of Corr. [r]	Gene	Std Dev	Coeff. of Corr. [r]	Gene	Std Dev	Coeff. of Corr. [r]
1	POP4	0.12	0.252	EIF3K	0.46	0.001	EIF3K	0.36	0.402	EIF3K	0.46	0.377
2	MORF4L1	0.15	0.001	RPLPO	0.71	0.001	PPIA	0.42	0.601	SDHA	0.72	0.333
3	HMBS	0.18	0.198	UXT1	0.72	0.361	POP4	0.51	0.530	PPIA	0.76	0.427
4	PPIB	0.20	0.626	EIF3M	0.73	0.251	SDHA	0.57	0.539	POP4	0.76	0.554
5	YWHAZ	0.22	0.221	SDHA	0.78	0.035	EIF3M	0.58	0.396	HMBS	0.81	0.173
6	EIF3M	0.32	0.435	HPRT1	0.81	0.321	GAPDH	0.61	0.397	PPIB	0.82	0.093
7	HNRNPL	0.34	0.001	PPIB	0.85	0.029	ACTB	0.61	0.125	YWHAZ	0.83	0.214
8	EIF3K	0.36	0.506	PPIA	0.86	0.138	YWHAZ	0.64	0.001	UXT1	0.89	0.001
9	SDHA	0.38	0.776	POP4	0.94	0.605	UXT1	0.65	0.095	HPRT1	0.91	0.384
10	UXT1	0.41	0.862	HMBS	0.95	0.363	HNRNPL	0.67	0.394	GAPDH	0.92	0.484
11	HPRT1	0.42	0.769	B2M	0.95	0.001	HMBS	0.70	0.720	EIF3M	0.94	0.579
12	RPLPO	0.43	0.763	MORF4L1	0.96	0.708	HPRT1	0.72	0.001	HNRNPL	1.06	0.609
13	GAPDH	0.44	0.440	YWHAZ	0.99	0.483	PPIB	0.79	0.288	MORF4L1	1.26	0.512
14	PPIA	0.53	0.681	GAPDH	1.26	0.722	ARF1	0.88	0.547	ARF1	1.31	0.434
15	ACTB	0.73	0.001	ACTB	1.34	0.444	B2M	1.16	0.202	ACTB	1.44	0.001
16	ARF1	0.82	0.506	HNRNPL	1.37	0.672	MORF4L1	1.19	0.124	B2M	1.78	0.485
17	B2M	0.85	0.689	ARF1	1.72	0.438	RPLPO	2.83	0.455	RPLPO	1.974	0.335

). The cut-off value for this stability measure is 1, below which all the genes are deemed to be stably expressed [9]; this was the case for all the genes investigated in stage 1 of our study. However, four and three out of the total of 17 HKGs displayed instability in stage 2 and stage 3, respectively, with stability values greater than 1. Furthermore, when every phase was considered, six HKGs were unstable.

3.4.3. Expression Stability of the HKGs by NormFinder Analysis

Expression stability values determined by the NormFinder algorithm are shown in

Table 4. Expression stability of HKGs by NormFinder analysis.

	Stage1		Stage2		Stage3		Total
Rank	Gene	Stability Value	Gene	Stability Value	Gene	Stability Value	Gene	Stability Value
1	PPIB	0.181	EIF3K	0.726	EIF3K	0.408	EIF3K	0.527
2	POP4	0.227	UXT1	0.817	PPIA	0.449	POP4	0.845
3	YWHAZ	0.266	EIF3M	1.077	POP4	0.523	SDHA	0.893
4	HMBS	0.292	MORF4L1	1.106	SDHA	0.591	PPIA	0.989
5	HPRT1	0.306	HPRT1	1.200	HMBS	0.623	EIF3M	1.017
6	MORF4L1	0.323	HMBS	1.322	GAPDH	0.643	PPIB	1.075
7	SDHA	0.372	POP4	1.327	EIF3M	0.658	HMBS	1.094
8	EIF3M	0.428	PPIB	1.328	HNRNPL	0.722	HNRNPL	1.122
9	EIF3K	0.429	SDHA	1.382	ACTB	0.823	GAPDH	1.145
10	RPLPO	0.454	PPIA	1.577	UXT1	0.871	UXT1	1.208
11	GAPDH	0.473	B2M	1.585	YWHAZ	0.924	HPRT1	1.296
12	HNRNPL	0.557	HNRNPL	1.678	PPIB	0.947	YWHAZ	1.306
13	UXT1	0.707	GAPDH	1.694	ARF1	1.012	MORF4L1	1.326
14	PPIA	0.745	YWHAZ	1.882	HPRT1	1.139	ARF1	1.713
15	ARF1	0.909	RPLPO	1.895	B2M	1.512	ACTB	1.869
16	B2M	1.11	ACTB	1.914	MORF4L1	1.573	B2M	2.139
17	ACTB	1.125	ARF1	2.587	RPLPO	3.255	RPLPO	2.635

. In stage 1, PPIB, selected from RNA-seq data, once again showed the best stability; this is the same as the result presented in the Delta Ct method. In stage 2 and stage 3, EIF3K was the most stable, while ARF1 and RPLPO were the least stable genes, respectively. In all samples, EIF3K, POP4, SDHA, and PPIA were the most stable, while ARF1, ACTB, B2M, and RPLPO were the least.

3.4.4. Expression Stability of the HKGs by geNorm Analysis

The geNorm algorithm calculated the M Value to determine the gene stability. The genes with lowest M value have the most stable expression and stepwise exclusion of genes with highest M values will result in a combination of two genes with the highest stability [7]. The M value of each gene in every stage and all samples are shown in

Table 5. Expression stability of HKGs by geNorm analysis.

	Stage1		Stage2		Stage3		Total
Rank	Gene	Stability Value	Gene	Stability Value	Gene	Stability Value	Gene	Stability Value
1	HMBS|POP4	0.070	EIF3K|UXT1	0.638	HMBS|SDHA	0.648	EIF3K|SDHA	0.940
2	PPIB	0.236	PPIB	0.942	PPIA	0.687	PPIA	1.006
3	MORF4L1	0.304	SDHA	1.245	EIF3K	0.711	HMBS	1.067
4	YWHAZ	0.327	HMBS	1.347	HNRNPL	0.779	POP4	1.126
5	HNRNPL	0.361	PPIA	1.393	POP4	0.824	PPIB	1.178
6	EIF3K	0.395	EIF3M	1.440	EIF3M	0.855	EIF3M	1.237
7	HPRT1	0.423	POP4	1.500	GAPDH	0.889	UXT1	1.306
8	RPLPO	0.445	MORF4L1	1.537	ACTB	0.936	HNRNPL	1.368
9	EIF3M	0.468	B2M	1.590	YWHAZ	0.967	GAPDH	1.416
10	SDHA	0.496	HPRT1	1.649	UXT1	0.992	MORF4L1	1.459
11	GAPDH	0.522	HNRNPL	1.737	PPIB	1.024	YWHAZ	1.495
12	UXT1	0.561	GAPDH	1.809	ARF1	1.057	HPRT1	1.533
13	PPIA	0.598	YWHAZ	1.886	HPRT1	1.099	ARF1	1.601
14	ARF1	0.640	ACTB	1.937	MORF4L1	1.169	ACTB	1.661
15	B2M	0.705	RPLPO	1.988	B2M	1.232	B2M	1.758
16	ACTB	0.766	ARF1	2.100	RPLPO	1.482	RPLPO	1.893

. In all samples, EIF3K and SDHA, with the M value of 0.940, were predicted to be the most stable genes, while B2M and RPLPO were found to be least stable, with M values of 1.758 and 1.893, respectively.

3.5. A Comprehensive Ranking of Four Methods

ReFinder was used to obtain an overall ranking of the selected 17 HKGs among the four methods. The result of the ReFinder algorithm are shown in

Table 6. Comprehensive rankings calculated using the the ReFinder method.

	Stage1		Stage2		Stage3		Total
Rank	Gene	Score	Gene	Score	Gene	STDEV	Gene	STDEV
1	POP4	1.41	EIF3K	1.00	EIF3K	1.41	EIF3K	1.00
2	PPIB	1.86	UXT1	1.68	PPIA	2.21	SDHA	2.45
3	HMBS	2.63	EIF3M	3.98	SDHA	2.99	POP4	3.16
4	YWHAZ	3.87	PPIB	4.74	POP4	3.57	PPIA	4.43
5	MORF4L1	3.94	MORF4L1	5.83	HMBS	4.26	PPIB	4.56
6	HPRT1	7.00	HPRT1	6.37	GAPDH	5.83	HMBS	6.09
7	EIF3K	7.71	HMBS	6.82	EIF3M	6.44	EIF3M	6.62
8	EIF3M	8.11	SDHA	7.14	HNRNPL	7.52	UXT1	7.00
9	SDHA	8.63	POP4	8.63	ACTB	8.45	GAPDH	8.97
10	HNRNPL	8.82	PPIA	8.80	YWHAZ	9.69	HNRNPL	9.39
11	RPLPO	9.97	B2M	9.59	UXT1	10.22	YWHAZ	9.87
12	GAPDH	11.49	RPLPO	11.77	PPIB	12.24	HPRT1	11.68
13	UXT1	13.24	HNRNPL	13.16	ARF1	13.24	MORF4L1	12.45
14	PPIA	13.74	GAPDH	13.21	HPRT1	13.47	ARF1	14.24
15	ARF1	15.24	YWHAZ	13.74	B2M	15.24	ACTB	14.47
16	B2M	16.24	ACTB	14.98	MORF4L1	15.74	B2M	16.00
17	ACTB	16.48	ARF1	17.00	RPLPO	17.00	RPLPO	17.00

. EIF3K, SDHA, POP4, and PPIA were the four most stable genes, while ARF1, MORF4L1, HPRT1, and YWHAZ were the four least stable genes across all samples.

3.6. Validation the Stability of HKGs with Target Genes

To validate the HKGs stability results, the most four stable genes (EIF3K, POP4, PPIA, and SDHA) and the least four stable genes (ACTB, B2M, RPLPO, and ARF1) were used to normalize the same target gene. When using the four most stable genes for normalization, the results showed that PAG8 and VEGFA were expressed in uterus caruncle tissue with the lowest level at stage 1, and that they increased in stage 2, and increased significantly more in stage 3 (p < 0.01), where they reached the highest level (Figure 4a,c). SPP1, whose expression reached the highest level in stage 2, expressed the least in stage 1 and the middle in stage 3 (Figure 4e). When it comes to the four least stable genes, there were hardly any significant differences between the stages (Figure 4b,d,f). In addition, different expression trends appeared among VEGFA and SPP1 when normalized to ACTB and B2M, respectively (Figure 4d,f).

A correlation analysis on the relative expression data of the four most and least stable candidate reference genes, normalized to three target genes, were performed. As shown in Figure 5, POP4 and PPIA had a high correlation coefficient with an R value of 0.97. It suggested that they have extremely similar normalization capabilities. However, the R value of ACTB and RPLPO is low i.e., 0.11; this demonstrates that they have hardly normalization capabilities.

4. Discussion

The uterus is an important organ in the reproductive system of most female mammals; it accommodates the embryonic and fetal development of one or more embryos until birth. During mammalian pregnancy, the maternal endometrium maintains a decidua reaction, including endometrial thickening, increase of blood vessels, hypersecretion of uterine glands, and stromal cell hypertrophy. These changes in morphology and composition in the maternal uterus suggest that a genetic regulation arises from transcriptional expression. However, the HKGs that normalize the expression of target genes have not been identified during pregnancy, especially in domestic animals. Thus, it is essential to explore suitable housekeeping genes. In this study, we refer to RNA-seq data to select novel candidate HKGs of goat uterus caruncle tissue, and qRT-PCR data were used for software analysis to obtain the most suitable HKGs. Our results showed that EIF3K, POP4, PPIA, and SDHA were the most stable HKGs in the uterus regardless of pregnant or nonpregnant state.

In order to obtain the stable HKGs from RNA-seq and qRT-PCR data, we set three parameters and used four methods to rigorously screen HKGs in uterine tissues. In reference to previous studies [12,18,19], the FPKM value, the coefficient of variation, and the maximum fold change were considered in order to measure the HKGs in the RNA-seq data. Meanwhile, geNorm, NormFinder, Bestkeeper, and Delta Ct methods were used to evaluate the stability of HKGs in qRT-PCR data; specifically, a total of 22 HKGs for qRT-PCR analysis. However, only 17 HKGs were used for further study. The reason we eliminated the five HKGs is based on the notion that a Ct of less than 29 is indicative for abundant gene expression [23], which is desirable for qRT-PCR analysis. Considering the accuracy and validity of the experiment, we set up a stricter Ct value line, viz. 28.

In our results, PPIB, a HKG we selected from RNA-seq data, showed good stability in stage 1 referring to nonpregnant period. PPIB is a kind of protein coding gene associated with secretory pathway. The reason why PPIB demonstrated better stability than the other 16 HKGs in uterus caruncles may relate to the tissue structure of the endometrium. The uterus caruncle is the pitting structure of endometrium, which contains an abundance of epithelial cells. We predict that epithelial cells probably contribute to the secretory process. A previous study [24] has also shown that PPIB demonstrated the best stability in LPS-stimulated and non-stimulated THP-1 and K562, which partly verified the foregoing prediction. Among the studies of HKGs performed on goats, we found a study [17] on Jianyang big-eared goats that showed that PPIB was the most stable gene. However, PPIB failed to show the best stability in the following stages. In stage 2 and stage 3, the fertilized ovum lodges in the uterus, establishing and maintaining pregnancy. A series of the aforementioned events might be more connected to genes that pertain to angiogenesis.

When it comes to samples in total stages, including stage 1, stage 2, and stage 3, EIF3K demonstrated the best stability. EIF3K (Eukaryotic translation initiation factor 3) has been suggested as one of the most stably expressed HKGs in bovine endometrium [25]. The uterus caruncle is the pitting structure of endometrium, so our results may partly validate previous research. In addition, it has been suggested that EIF3K is one of the most stably expressed HKGs in bovine endometrium, and is also stably expressed in bovine adipose tissue, mammary gland, muscle, liver, and goat adipose tissue [20,26]. To date, no publications were found on EIF3K as an unstable HKG. This may indicate the potential of EIF3K to serve as a stable HKG in goat uterus tissue.

It is interesting that traditional HKGs such as ACTB and GAPDH did not perform well in our research. ACTB was one of the four least stable genes, while GAPDH showed the medium stability overall. It seems to be a common conclusion that classical reference genes did not have good stability in the studies of HKGs [24,27,28,29,30]. This finding reminds us of the need to accurately validate candidate internal control genes in the study before use in gene expression studies using qRT-PCR.

When it comes to methodology in this study, we combine RNA-seq data with references to select HKGs following the use of qRT-PCR experiments to validate the selected HKGs; subsequently, the target genes of uterus caruncle tissue were used for further validation. We used four algorithms (geNorm, Delta Ct, NormFinder, and BestKeeper) to calculate the stability of HKGs. Then, ReFinder was applied to evaluate the comprehensive stability of HKGs. In general, we used all the latest methods for screening HKGs, which leads us to a relatively reliable result.

For the number of HKGs for normalization, we only chose single genes instead of gene combinations. This was determined by geNorm based on the average pairwise variation (V_n/V_n+1) value. As a routine, 0.15 was proposed as the cut-off value for the V value [7], below which the inclusion of an additional control gene was not required. However, all the V_n/_n+1 values in our study were less than 0.15 (Table S2), indicating that increasing the number of genes did not significantly change the value of the normalization analyses. Therefore, a single top stable HKG might be a good choice for qRT-PCR assay in our study.

Previous studies have shown that SPP1, VEGFA, and PAG8 [31,32,33] genes were the marker genes for uterus and placenta tissues. In addition, they were differentially expressed in our RNA-seq data (data was not shown) in the uterus caruncle tissues. Therefore, SPP1, VEGFA, and PAG8 genes were selected as target genes for normalization. SPP1 plays an important role in the regulation of maternal-fetus angiogenesis and calcium balance after implantation [34,35]. VEGFA (vascular endothelial growth factor A) has strong pro-angiogenic effect on embryonic vasculogenesis [36]. Early expression of PAG genes, including PAG8, in the conceptus may contribute to successful development of early pregnancy and possibly alter the mechanisms related to embryo survival, such as prostaglandin synthesis [33]. In our study, the extremely upregulated PAG8 and VEGFA genes during pregnancy were shown to be significantly normalized to the most stable HKGs (EIF3K, POP4, PPIA, and SDHA). However, hardly any significance, and even severe variance, appeared when they were normalized to the four least stable HKGs (ACTB, B2M, RPLPO, and ARF1). In addition, the expression of SPP1 demonstrated some variation between the most stable and unstable HKGs. The suitable HKGs should be carefully chosen considering the precision of qRT-PCR, as unstable reference genes will result in incorrect interpretation of expression data.

5. Conclusions

In our study, four different algorithms and one comprehensive index were used to analyze a panel of HKGs. All four algorithms gave consistent results. PPIB and EIF3K, POP4, PPIA, and SDHA proved to be the most suitable HKGs in non-pregnant stage and both non-pregnant and pregnancy stages, respectively. ACTB, B2M, RPLPO, and ARF1 were the four least stable HKGs, suggesting they are not suitable for expression data normalization in studies of goat uterus caruncle tissue.